Feature Factory¶

The product-organisation anti-pattern of measuring success by features shipped per unit time while never asking whether any feature changed an outcome — a proxy output displacing the target it was meant to track, Goodhart's Law in the product operating loop.

Version: v4 · 2026-09-28 · History
Domain-specific #: 392
Domain group: Professional & Organizational Practice
Origin domain: Organizational & Management Science
Subdomain: Product Development → Organizational & Management Science

Core Idea¶

A feature factory is the product-organisation anti-pattern in which success is measured by the number of features shipped per unit time, the operating rhythm is calibrated to keep throughput high, and whether any shipped feature changed user behaviour, revenue, or retention goes unasked. The output metric substitutes for the outcome metric through a four-stage move: a target outcome motivates the org, a measurable proxy output is rewarded, the loop closes on the proxy, and output and outcome decouple — Goodhart's Law in the product loop.

How would you explain it like I'm…

The Cake-Counting Bakery

Imagine a bakery that only counts how many cakes it makes each day and never checks if anyone eats them or likes them. They make more and more cakes and feel busy and proud, but nobody knows if it helped. A team that builds things that way is called a Feature Factory.

Counting Features, Not Results

A Feature Factory is a company team that judges itself by how many new features it finishes, not by whether those features help anybody. The real goal might be keeping customers happy or earning money. But counting finished features is easier, so the team starts chasing that number instead. Their charts look great, while the real goals might not improve at all. Since nothing punishes useless features, teams can even split work into tiny pieces just to look busier.

Outputs Over Outcomes Trap

A Feature Factory is a product-team anti-pattern where success is measured by how many features get shipped, and nobody seriously checks whether those features changed user behavior, revenue, or retention. The output number (features, story points, sprints) replaces the outcome the work was supposed to serve. It develops in stages: a real goal justifies the team; an output measure is easy to count and control; the team starts rewarding the output as if it were the goal; and the two drift apart because shipping useless features costs nothing. Gaming follows, with smaller features and more launches. John Cutler popularized the term, and it's an example of Goodhart's Law: when a measure becomes a target, it stops tracking what it was meant to.

The Feature Factory is an organizational anti-pattern in which a product organization measures success by features shipped per unit time, tunes its operating rhythm for throughput, and leaves unasked, or asks only ceremonially, whether shipped features moved the outcomes they were meant to serve. Output metrics such as features delivered, story points, and sprints closed substitute for outcome metrics such as retention, revenue, or mission impact. The mechanism is a four-stage substitution: a target outcome motivates the organization; a proxy output is measurable, controllable, and locally produced; the operating loop rewards and reports the proxy as if it were the target; and output and outcome decouple, because nothing penalizes outcome-neutral shipping. Gaming then follows, with smaller features, more sprints, and more launches, while dashboards look productive and real metrics drift independently. John Cutler codified the term within the outcomes-versus-outputs literature. Structurally, it is Goodhart's Law applied to the product operating loop.

Scope of Application¶

Feature factory lives within product organisations, across every surface form the proxy-for-outcome substitution takes in software-product and digital-delivery work.

Roadmap-by-feature-list planning — a feature list carrying no outcome hypotheses or kill criteria.
Velocity-as-headline-metric cultures — story throughput as team health, outcomes never gating the next sprint.
Launch-then-move-on practice — features shipped and left behind with no instrumented validation.
Sales-led roadmaps — every feature promised to a deal, no one revisiting whether it is used.
Innovation theatre and after-the-fact bar inflation — feature counts substituting for outcome movement.

Clarity¶

Naming the feature factory reframes "we ship a lot but the numbers aren't moving" from an execution story into a metric-substitution diagnosis: the team may be correctly optimising the metric placed in front of it, and the defect is which metric the loop closes on. It sharpens the output/outcome distinction throughput charts erase and licenses one diagnostic question: for each shipped feature, what outcome was hypothesised, and was it measured?

Manages Complexity¶

Product organisations present a scatter of dysfunctions — sales-led roadmaps, velocity obsession, transformation theatre — that a leader might fix separately. Feature factory collapses them into one mechanism: the loop closed on a proxy. The diagnosis reduces to a single measurable ratio (the answer-rate to the outcome-hypothesis question), and one intervention family fixes all variants: pair every committed output with an outcome hypothesis and kill criterion before commit.

Abstract Reasoning¶

The proxy-target decoupling licenses diagnosis from the answer-rate not the throughput (and reading gaming tells), interventionist rewiring of the operating-metric graph rather than exhorting the team (predicting throughput should fall when fixed), boundary-drawing (is the outcome knowable and abdicated, versus deliberately deferred exploratory work), and an order-of-events forecast (the decoupling widens over time and surfaces late).

Knowledge Transfer¶

Within product organisations the diagnosis transfers as mechanism across every surface form — one mechanism with the pathology swapped — the answer-rate diagnostic and forcing-function fix carrying intact, supported by the outcomes-vs-outputs literature (Cutler, Cagan, Torres). The named concept stops at product orgs; its breadth is one substrate replayed. What travels to distant systems is the target_measure_substitution / Goodhart parent (test-score gaming, hospital waiting times, command-economy tonnage); the product-management vocabulary and the name stay home.

Relationships to Other Abstractions¶

Current abstraction Feature Factory Domain-specific

Parents (2) — outgoing typed relations

Feature Factory is a kind of, typical Build Trap Domain-specific

A Feature Factory is the characteristic request-and-throughput realization of the broader Build Trap, though the full three-dial regime is not guaranteed.
Feature Factory is a decomposition of Proxy-Target Divergence Prime

Stripped of product vocabulary, the factory is an operating apparatus closed on a controllable proxy after that proxy has decoupled from its target.

Hierarchy paths (2) — routes to 2 parentless roots

Feature Factory → is a kind of, typical Build Trap

Show alternative path (1)

Neighborhood in Abstraction Space¶

Feature Factory sits in a moderately populated region (58^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25